Multi-Aspect Explainable Inductive Relation Prediction by Sentence Transformer

Zhixiang Su, Di Wang, Chunyan Miao, Lizhen Cui
1/4/2023
cs.CLcs.LG

No official code found. Best implementation: ZhixiangSu/KRST (7 stars, last commit 2024-02-08).

Abstract

Recent studies on knowledge graphs (KGs) show that path-based methods empowered by pre-trained language models perform well in the provision of inductive and explainable relation predictions. In this paper, we introduce the concepts of relation path coverage and relation path confidence to filter out unreliable paths prior to model training to elevate the model performance. Moreover, we propose Knowledge Reasoning Sentence Transformer (KRST) to predict inductive relations in KGs. KRST is designed to encode the extracted reliable paths in KGs, allowing us to properly cluster paths and provide multi-aspect explanations. We conduct extensive experiments on three real-world datasets. The experimental results show that compared to SOTA models, KRST achieves the best performance in most transductive and inductive test cases (4 of 6), and in 11 of 12 few-shot test cases.

View on arXivView PDF

Code Implementations(1)

ZhixiangSu/KRSTCommunity100%
71Shell, PythonAug 12, 20222 years ago

Similar Papers

Inductive Relation Prediction from Relational Paths and Context with Hierarchical Transformers

88% match1 repo72023

Replacing Paths with Connection-Biased Attention for Knowledge Graph Completion

84% match1 repo2024

Relations Prediction for Knowledge Graph Completion using Large Language Models

83% match2024

Knowledge Graph Completion using Structural and Textual Embeddings

82% match1 repo42024

A Benchmark for Semi-Inductive Link Prediction in Knowledge Graphs

82% match1 repo52023

Knowledge Reasoning Language Model: Unifying Knowledge and Language for Inductive Knowledge Graph Reasoning

81% match2025

Cite this paper

@article{su2023multiaspect,
  title  = {Multi-Aspect Explainable Inductive Relation Prediction by Sentence Transformer},
  author = {Zhixiang Su and Di Wang and Chunyan Miao and Lizhen Cui},
  year   = {2023},
  eprint = {2301.01664},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2301.01664v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2301.01664.svg)](https://codeofpaper.com/paper/2301.01664)

Discussion